Forecasting the output power of photovoltaic (PV) arrays using 3 & 6 years of weather and PV output power data using both machine & deep learning
Python version 3.10.8 was used in this study.
"Further Consolidated Data, HnL": The Dataset. It contains the datasets for each city for 3 and 6 years in length. The first five sheets has 3 years of data for all five cities, while the last four sheets have the complete 6 years of data for the cities in California. The units of the 22 parameters were omitted from the dataset for simplcity when testing, but the parameters and their units can be found in the paper.
"Main_nn": The main function for the deep learning models developed. It relies on "HelperFunctions", "MineLSTM", "MineGRU". To adjust which city the models are trained or tested on, simply adjust the int that correlates to a different key value of the dict of cities. Only sequential splitting is conducted.
"Testing_Algorithmns" The works as the main for the machine learning models. Relies on "HelperFunctions". While sequential or random splitting of the datatset is possible, it is currently configured for random splitting.
"MineLSTM" & "MineGRU": The two neural networks that were adapted for regression analyses on both 3 and 6 years of data across 5 different cities.
"Graphing_Thesis_Results": A supplemenary file that strictly graphs the data generated by the other files. Note, the file does not receive the results automatically, it is relient on manually pasting in the numerical results generated.